Research Article

Prostate Cancer Cells with Stem Cell Characteristics

Reconstitute the Original Human Tumor In vivo
1

1,3

Guangyu Gu, Jialing Yuan, Marcia Wills, and Susan Kasper

1

Department of Urologic Surgery, Vanderbilt University Medical Center; 2Department of Pathology, Vanderbilt
Childrens Hospital; and 3The Vanderbilt-Ingram Cancer Center, Nashville, Tennessee

Abstract
Cancer may arise from a cancer stem/progenitor cell that
shares characteristics with its normal counterpart. We report
the reconstitution of the original human prostate cancer
specimen from epithelial cell lines (termed HPET for human
prostate epithelial/hTERT) derived from this sample. These
tumors can be described in terms of Gleason score, a
classification not applied to any of the transgenic mouse
models currently developed to mimic human disease. Immunohistochemical and Western blot analyses indicate that they
do not express androgen receptor or p63, similar to that
reported for prostate stem cells. These cell lines also express
embryonic stem markers (Oct4, Nanog, and Sox2) as well as
early progenitor cell markers (CD44 and Nestin) in vitro.
Clonally derived HPET cells reconstitute the original human
tumor in vivo and differentiate into the three prostate
epithelial cell lineages, indicating that they arise from a
common stem/progenitor cell. Serial transplantation experiments reconstitute the tumors, suggesting that a fraction of
parental or clonally derived HPET cells have self-renewal
potential. Thus, this model may enhance our understanding of
human tumor development and provide a mechanism for
studying cancer stem/progenitor cells in differentiation,
tumorigenesis, preclinical testing, and the development of
drug resistance. [Cancer Res 2007;67(10):480715]

Introduction
Cancer stem cells are thought to arise from a multipotent stem
cell that has accumulated genetic alterations during its long life
span (1). An alternative hypothesis suggests that normal stem cell
differentiation may be arrested during cell determination and
differentiation, thereby affecting the development of malignant
potential (2). Cancer stem/progenitor cells may exhibit characteristics similar to normal stem cells. For example, stem cells express
the protein telomerase, which maintains telomere length and
facilitates continuous cell division (3). They divide and give rise to
multiple progenitor cell types that down-regulate telomerase
activity during gestation (4) and early cell passage in vitro (5).
Thus, loss of telomerase activity seems to limit the mitotic capacity
of progenitor cells. Roy et al. showed that hTERT-immortalized
progenitor cells from human fetal spinal cord sustained telomerase

activity and continued to give rise to mature spinal interneurons

and motor neurons, indicating that hTERT overexpression permitted generation of progenitor lines able to give rise to phenotypically
restricted neurons (3). hTERT has also been successfully used to
immortalize human prostate epithelial (HPE) cells from primary
tumors to generate RC-58T/hTERT and 957E/hTERT cell lines (6, 7).
Recently, Litvinov et al. reported that the 957E/hTERT cell line
contained prostate stem cells (8). Thus, although immortalization
using hTERT is not equivalent to cancer stem cell isolation, it can be
used to extend the mitotic capacity of progenitor cells and
immortalize cells with stem cell characteristics.
Cancer stem cells share similar properties of self-renewal and
differentiation as well as a similar phenotype with adult stem
cells isolated from the same tissue (1, 2). The most compelling
evidence of the existence of prostate stem cells in the basal cell
compartment is derived from the mouse castration model where
androgen withdrawal results in glandular involution and apoptosis in f90% of epithelial cells but leaves the basal cell layer intact
(9). Androgen replacement restores basal and luminal epithelial
cells, inducing proliferation and differentiation (9). Slow-cycling
cells retaining bromodeoxyuridine labeling following androgen
withdrawal/replacement experiments have been identified in both
basal and luminal cell compartments, implying that prostate stem
cells are not restricted to one epithelial cell compartment (10).
However, direct evidence of a putative prostate cancer stem cell
has not been reported.
We routinely culture primary HPE cells to study the role of the
androgen receptor (AR) in the development of androgen-independent prostate cancer (11). To facilitate these studies, we attempted
to derive HPET (where T represents hTERT) cell lines through the
stable integration of hTERT into these primary cells. Here, we
report the characterization of these cell lines. Surprisingly, they
exhibit stem cell characteristics, expressing embryonic stem cell
markers, including Oct4, Nanog, and Sox-2, in addition to the early
progenitor cell markers CD133, CD44, and nestin. HPET cells do
not express p63 and AR, similar to other reports on prostate stem
cells (6, 7). Most importantly, clonally derived HPET cells are
capable of reconstituting the original prostate tumor from which
they were derived and retain the ability to differentiate into basal,
luminal, and neuroendocrine epithelial cell types of the prostate
in vivo. Therefore, we present the first direct evidence for the
existence of a putative prostate cancer stem cell.

Materials and Methods

Note: Supplementary data for this article are available at Cancer Research Online
(http://cancerres.aacrjournals.org/).
Requests for reprints: Susan Kasper, Department of Urologic Surgery, Vanderbilt
University Medical Center, A-1302 Medical Center North, 1161 21st Avenue South,
Nashville, TN 37232-2765. Phone: 615-343-5921; Fax: 615-322-8990; E-mail:
susan.kasper@vanderbilt.edu.
I2007 American Association for Cancer Research.
doi:10.1158/0008-5472.CAN-06-4608

www.aacrjournals.org

hTERT construct and Lentiviral particles. The 3.5-kb EcoRI-SalI

fragment from pBABE-PURO-hTERT (provided by Dr. Judith Campisi,
Lawrence Berkeley National Laboratory, Berkeley, CA) was subcloned into
the EcoRI and SalI sites of pLenti-EGFP (Invitrogen/Life Technologies) to
create pLenti-hTERT-EGFP. Viral particles were produced by transfecting
293 FT cells with 3 Ag pLenti-EGFP expression plasmid DNA and 9 Ag
ViraPower packaging mix (Invitrogen/Life Technologies) per 10-cm plate.

4807

Cancer Res 2007; 67: (10). May 15, 2007

Cancer Research
Supernatants were collected 48 h after transfection, filter sterilized, and
stored at 80jC. All procedures were done as recommended by Vanderbilt
University safety regulations for Lentivirus usage.
Establishment of HPET cell lines. Radical retropubic prostatectomy
specimens were obtained in compliance with the laws and institutional
guidelines approved by the Institutional Review Board Committee of
Vanderbilt University. Prostate tissues were processed, and HPE cells were
cultured and maintained as described previously (12). Briefly, punch biopsies
of prostate tissue were minced; tissue fragments were plated on Primaria
plates (Falcon); and HPE cells were cultured in epithelial cellselective
medium (RPMI 1640 supplemented with 2.5% charcoal-stripped, heatinactivated fetal bovine serum; 20 mmol/L HEPES buffer; 100 units/mL
penicillin; 100 mg/mL streptomycin; 0.25 mg/mL amphotericin B; 50 mg/mL
gentamicin; 56 mg/mL bovine pituitary extract; 1 insulin-transferrinselenium; 10 ng/mL epidermal growth factor; and 50 ng/mL cholera toxin)
and maintained in 5% CO2 at 37jC.
The parental HPET cell line was established by transducing HPE cells
(passage 2) with supernatant containing pLenti-hTERT-EGFP particles and
culturing them in selection medium containing 10 Ag/mL blasticidin. The
HPET-1, HPET-5, HPET-11, and HPET-13 clonal lines were established by
serial dilution and verified visually by at least two individuals. Telomerase
activity was determined as described (see Fig. 1C; ref. 13).
Anchorage-independent growth assays. Clonogenicity generation
assay were done as previously described with minor modifications (14).
Briefly, subconfluent cultures were harvested and plated in triplicate 35-mm
wells of six-well plates in 2 mL medium with 0.5% low gelling temperature
agarose at a concentration of 5  103 cells per mL on a 2-mL base layer of
1.0% agarose in medium. Cells were fed additional medium weekly. After
4 weeks, colonies with >40 cells were counted under an optical microscope.
For spheroid generation assay, cells were plated on ultralow adhesion tissue
culture dishes (Corning) at the density of 2  104 per mL in epithelial cell
selective medium without serum. Pictures were taken at 3 days after plating.
Cell differentiation on Matrigel. HPET and HPET1 cells were
trypsinized, and 2.5  105 cells were mixed in 50 AL of 50% (v/v) epithelial
medium and Matrigel (BD Biosciences; ref. 15). The mixture solidified in a
six-well Petri dish at 37jC and then detached from the dish. Four milliliters
of epithelial medium were added.
Immunofluorescence microscopy. Primary cultured cells were harvested by trypsinization and attached onto glass slides (VWR) by incubation
at 37jC overnight. Slides were fixed in 4% paraformaldehyde in PBS for
15 min, washed thrice in PBS, and permeabilized for 5 min on ice in PBS +
0.1% Triton X-100. Slides were blocked in PBS + 3% bovine serum albumin +
3% donkey serum for 30 min at room temperature. The slides were incubated
with primary antibodies for 1 h at room temperature and subsequently
incubated with Alexa Fluor 594 and/or Alexa Fluor 488conjugated
secondary antibody (Molecular Probes) for 1 h at room temperature. Then,
cells were washed in PBS and mounted using Vectashield mounting medium
containing 4,6-diamidino-2-phenylindole to counterstain nuclei (Vector
Laboratories). Images were captured with a Leica fluorescence microscope
equipped with a digital camera (Vanderbilt University Medical Center
Immunohistochemistry Core Laboratory). To show specificity of staining, the
following controls were included: omission of either the primary antibody or
the secondary antibodies.
Reverse transcription-PCR. Total cellular RNA from HPET, HPET-1,
HPET-5, HPET-11, HPET-13, and LNCaP cell lines was extracted using TRI
Reagent (Sigma) according to the manufacturers instructions. Two micrograms of total RNA for all cell lines were used to generate cDNA (RT-PCR
Access System, Promega). cDNA was amplified using 1 AL of the reverse
transcriptase reaction products in 25 AL with 10 pmol of the primers for
35 cycles. The PCR products were analyzed by electrophoresis on 1.5%
agarose gels. Primers are listed in Supplementary Table S1. Glyceraldehyde3-phosphate dehydrogenase was used as the internal control in all reactions.
Western blot analysis. Cells were extracted with ice-cold 1 lysis buffer
(Promega) and a cocktail of protease and phosphatase inhibitors (Sigma).
Lysates were then centrifuged for 10 min at 14,000  g at 4jC; 20 Ag protein
from each supernatant was subjected to 4% to 20% SDS-PAGE and
transferred to polyvinylidene difluoride and blocked and probed overnight

Cancer Res 2007; 67: (10). May 15, 2007

at 4jC with primary antibody. Peroxidase-conjugated secondary antibody

was added at a 1:10,000 dilution and developed with enhanced
chemiluminescence.
Immunohistochemical analysis. Tissues were fixed in 10% buffered
formalin overnight followed by transfer to 50% alcohol. The paraffinembedded tissues were sectioned (5 Am). Sections were deparaffinized and
rehydrated in ethanol solutions. After antigen unmasking by boiling in 10
mmol/L sodium citrate buffer (pH 6.0) for 20 min, the sections were treated
with 3% hydrogen peroxide for 5 min. The following detection and
visualization procedures were done according to manufacturers protocol
(Vector Laboratories). AR, p63, green fluorescent protein (GFP), Ki-67, CD44,
Nestin (Santa Cruz), cytokeratin 8 (CK8), CK18, prostatic acid phosphatase
(PAP; Sigma), E-cadherin, Synaptophysin (BD Transduction Laboratories),
34hE12 (DAKO), and hMT (Chemicon) were used. Negative control slides
were done without primary antibodies.
Tissue recombination assay. All animals were housed in pathogen-free
units at Vanderbilt University Medical Center, and all procedures were done
in compliance with Institutional Animal Care and Use Committee
regulations. Rat urogenital sinus mesenchyme (rUGM) was prepared from
18-day embryonic fetuses of pregnant Sprague-Dawley rats (Harlan
Sprague-Dawley, Inc.). Dissection and separation of urogenital sinus
epithelium and UGM were done as previously described (16). HPET and
HPET-5 cells (1  105) were then recombined with rUGM (2  105) in
neutralized type I rat tail collagen gels. Tissue recombinants were grafted
beneath the renal capsule of adult homozygous severe combined
immunodeficient (SCID) male nude mice (Charles River Laboratories) as
described. Grafts were harvested at 12 weeks. The number of tumor grafts
examined was detailed in Supplementary Table S2.

Results
hTERT overexpression immortalizes HPE cells. To immortalize HPE cells, we cultured primary HPE cells from human Gleason
grade 8 to 9 prostate cancer specimens under conditions favoring
epithelial cell growth (12). HPE cells (passage 2) were transduced
with pLenti-hTERT-EGFP particles and cultured in selection
medium containing 10 Ag/mL blasticidin to establish the parental
hTERT-expressing (HPET) cell line. HPE cells expressing hTERT
and GFP (labeled as HPET cells) were already evident after 1 week
in selection medium containing 10 Ag/mL blasticidin (Fig. 1A and
B). They grew as adherent cells with epithelial cell morphology
(Fig. 1D). HPET cells have successfully been cultured over a 1-year
period with no evidence of decreased proliferative capacity. The
original HPE cell culture exhibited measurable telomerase activity,
and this activity was increased in hTERT-immortalized parental
HPET and clonal HPET-1, HPET-5, HPET-11, and HPET-13 cell lines
(Fig. 1C). The levels of hTERT activity in the parental HPET and
clonal cells lines seemed indistinguishable. Clonally derived HPET
cell lines were established by serial dilution. Six 96-well plates were
seeded at one to a few cells per well, and 55 wells contained single
cells. Despite increased hTERT expression, only four single cell
clones (labeled HPET-1, HPET-5, and HPET-11 and HPET-13)
survived and expanded to over one million cells. These clonally
derived cells exhibited typical cobblestone morphology seen in
epithelial cell cultures (Fig. 1D). All remaining clones and single
cells underwent senescence within 4 to 6 weeks.
Parental HPET and clonal HPET cells do not express AR or
p63. In agreement with previous reports on putative prostate
stem/progenitor cells (17, 18), parental and clonal HPET cells do
not express AR protein (Fig. 2A and B), or secretory proteins such
as prostate-specific antigen (PSA) and PAP (data not shown). AR is
not observed even when cells are cultured in the presence of
androgen (data not shown). Data obtained from HPET-5 cells are
representative of that observed in the HPET-1, HPET-5, HPET-11,

4808

www.aacrjournals.org

Prostate Cancer Cells with Stem Cell Characteristics

Figure 1. Generation of the HPET and HPET-5 cell lines. A, primary HPE cells infected with virus. B, confirmation of transduction using GFP expression.
C, telomerase activity in cell lysates from parental HPET cells; clonal HPET-1, HPET-5, HPET-11, and HPET-13 cells; and original non-immortalized HPE cells.
Cells were analyzed for telomerase activity by TRAP assay according to the manufacturers instructions (TRAPeze, Chemicon). Telomerase activity was assayed from
2 Ag of cell extract including telomerase-positive control cells provided in the TRAPeze kit. Lysis buffer alone and 2 Ag of protein extract derived from all samples
was heat treated at 85jC for 5 min (which inactivates telomerase) and were used as negative controls. TSR8 DNA template provided in the TRAPeze kit served as
a PCR-positive control. D, representative micrographs of established parental HPET and clonal HPET-1, HPET-5, HPET-11, and HPET-13 cell lines as indicated.
Bar, 50 Am.

and HPET-13 clonal lines. Less than 1% HPET cells express minimal
levels of the basal cell marker p63, whereas p63 is absent in the
clonal HPET cell lines (Fig. 2A). CK8 and CK18 are expressed at low
levels in a small number of HPET and HPET-5 cells, indicating
limited luminal differentiation in culture. Together, these findings
suggest that HPET and HPET-5 cells are not derived from a p63expressing basal cell.
In contrast to previous reports on putative mammary and
prostate stem cells (17), we do not observe Hoechst 33342 dye
exclusion in either parental or clonal HPET cells (data not shown).
This supports a recent report on the generation of a functional
mammary gland from a single stem cell (19), suggesting that
Hoechst 33342 dye exclusion is not representative of all stem/
progenitor cells. HPET and clonal HPET cells were subsequently
screened with a panel of stem cell markers to determine their
molecular profile.
Parental HPET and clonal HPET-5 cells express embryonic
stem cell and early progenitor cell markers and exhibit other
stem celllike properties. As anticipated, expression levels of
embryonic stem cell and prostate progenitor markers varied
between the parental and clonal HPET cell lines because they
were derived from different clones (Fig. 2BD). However, all lines
express mRNA for the embryonic stem cell markers Nanog, Oct4,
and Sox2 that regulate pluripotency and self-renewal in mouse

www.aacrjournals.org

and human embryonic stem cells (2022). Early prostate

progenitor cell markers, such as CD44 (23), CD133 (24, 25),
Nestin (26), and the receptor tyrosine kinase c-kit (27), are also
expressed (Fig. 2C). CD44 and Nestin expression in HPET and
HPET-5 cells was confirmed by immunocytochemical and Western
blot analysis (Fig. 2B and D). Thus, HPET and HPET-5 cells
express both stem cell and early progenitor cell markers.
To test the progenitor capacity of (AR )p63 CD44+Nestin+ HPET
and HPET-5 cells, we determined their ability to form spheroids
and reconstitute prostatic glandular structure in vitro. Typically,
both cell types form spheroids in low adherence culture (Fig. 3A).
HPET cells form large and small colonies and exhibit a 3-fold higher
colony frequency (Fig. 3C and D), indicating that the parental line
is not clonal. Clonal HPET-5 cells only form uniformly small
colonies. When cultured on Matrigel (Fig. 3B), both form glandularlike structures with lumens; however, the branched structures of
HPET-5 cells seem larger and more complex. These findings
indicate that both parental and clonal HPET cells possess stem
celllike characteristics.
HPET cells of single-cell origin recapitulate the original
human tumor with multipotent differentiation. To determine
whether HPET cells recapitulate the original prostate tumor in vivo,
they were recombined with inductive rat embryonic mesenchyme
and grafted under the renal capsule of male SCID mice (Fig. 4A

4809

Cancer Res 2007; 67: (10). May 15, 2007

Cancer Research

and B). HPET cells form tumors in 3 months with an average size of
6  6  3 mm. The human origin of the tumor is confirmed by GFP
staining and human-specific mitochondrial protein hMT (28)
expression (Fig. 4C). Importantly, histopathologic and immunohistochemical analyses indicate that HPET tumors are nearly identical
to the original human prostate cancer biopsy from which they were
derived (Fig. 5A and Fig. 6A). They show typical heterogeneity
observed in human prostate cancer and are classified as Gleason
4 + 4. Gleason 5 patterns like those observed in the original tumor
(Gleason 5 + 4) are also present (Fig. 6A).

Immunohistochemical analysis indicates that AR is reexpressed in most cells, although some tumor foci are AR ,
similar to that observed in advanced prostate adenocarcinomas
(29). HPET tumors also re-express the differentiated protein PAP,
although PSA was not observed (data not shown). All tumors
exhibit a high proliferative index as seen by Ki67 staining.
Significantly, the three prostate epithelial cell lineages are
represented (Fig. 5A and D). Luminal epithelial cells are identified
by their high levels of E-cadherin and CK8/18 expression. In
regions where cribriform high-grade prostatic intraepithelial

Figure 2. Characterization of AR,

p63, cytokeratin, and stem cell markers
in parental HPET and clonal HPET-5
cells. A, immunofluorescence staining
with anti-AR, anti-p63, and anti-CK8/18
antibodies specific to the human proteins.
The prostate cancer cell line LNCaP served
as a control. B, Western blot analysis
for AR, p63, Nestin, and CD44. Lane 1,
parental HPET cell line; lanes 2 to 5,
clonal cell lines HPET-1, HPET-5, HPET11, and HPET-13, respectively. LNCaP
cells (lane 6) are used as positive
controls except for p63 expression where
human benign prostatic hyperplasia-1
(BPH ) cells serve as positive control.
C, expression of embryonic stem cell and
early progenitor cell markers. Primers are
listed in Supplementary Table S1. Controls:
LNCaP cells (lane 6 ); no template in
the reverse transcription-PCR reaction
(lane 7). Glyceraldehyde-3-phosphate
dehydrogenase (GAPDH ) is the internal
control. D, CD44 and Nestin expression
by immunofluorescence staining with
anti-CD44 and anti-Nestin antibodies.
Bar, 50 Am.

Cancer Res 2007; 67: (10). May 15, 2007

4810

www.aacrjournals.org

Prostate Cancer Cells with Stem Cell Characteristics

Figure 3. Characterization
of the progenitor capacity of
(AR )p63 CD44+Nestin+ HPET and
HPET-5 cells in vitro. A, multicellular
spheroid formation on lowadherence culture plates. Bar, 50 Am.
B, colonies and branching structures of
(AR )p63 CD44+Nestin+ HPET and
HPET-5 cells on Matrigel. Representative
H&E-stained sections (bottom ). Bar ,
200 Am (top ) and 20 Am (bottom ).
C, colony-forming ability on agar. Bar,
100 Am. D, histogram of Fig. 2C. Columns,
mean (n = 3); bars, SE.

neoplasia are present, basal epithelial cells expressing p63 and

34hE12 are still observed surrounding the more glandular
structures in the tumor grafts as well as in the original human
tumor. Other cells express the neuroendocrine cell marker
synaptophysin similar to that observed in the original human
tumor. Thus, HPET cells recapitulate the histopathology of the
original prostate adenocarcinoma in vivo.
The observation that HPET cells are AR and p63 suggests
that they have stem celllike characteristics. Alternatively, they
could contain prostate cancer stem/progenitor cells. To pursue
this further, clonal lines were analyzed in the tissue recombination assay to determine whether they could reconstitute the
original tumor in vivo. Similar to parental HPET line, HPET
clones, derived from a single cell and represented by HPET-5,

recapitulate the histopathology of the original prostate adenocarcinoma in vivo and differentiate into the three prostate epithelial
cell lineages in vivo (Fig. 5B). Thus, it seems that these three cell
lineages arise from a common (AR )p63 CD44+Nestin+ stem/
progenitor cell.
Within most tumors, there likely exists a subpopulation of cancer
stem cells that continue to potentiate tumorigenesis. As seen in
Fig. 5C, CD44 expression occurs at the cell membrane in both basal
and luminal-like cells. Nestin expression is observed in a few cells
within the basal and luminal cell compartments, suggesting that
prostate cancer stem/progenitor cells are not restricted to the basal
cell compartment. These findings concur with the observation that
slow-cycling/stem cells are present in both basal and luminal cell
compartments (10). Serial transplantation experiments were done

Figure 4. Single-cell origin

(AR )p63 CD44+Nestin+ HPET-5 cells
can regenerate prostate glandular structure
in vivo. A, diagrammatic representation
of the tissue recombination assay. rat
eUGM, rat day 18 embryonic urogenital
mesenchyme. B, prostate tumor grafts
under the capsule of SCID mouse kidneys.
Bar, 2 mm. C, GFP and human
mitochondrial protein (hMT ) expression
in prostate epithelial cells. Number of tumor
grafts examined was detailed in
Supplementary Table S2. Bar , 20 Am.

www.aacrjournals.org

4811

Cancer Res 2007; 67: (10). May 15, 2007

Cancer Research

Figure 5. (AR )p63 CD44+Nestin+ HPET-5 cells differentiate into the epithelial cell lineages of the prostate and recapitulate the original human tumor (HT ). A, similar
to that of the original HT specimen, epithelial cells in HPET and HPET-5 tumors re-expressed AR and PAP. The high proliferative index in the human tumor
(as seen by antiKi-67 staining) was reproduced in HPET and HPET-5 tumors. Bar, 50 Am. B, epithelial cell differentiation in tumors derived from single-cell origin
(AR )p63 CD44+Nestin+ HPET-5 cells. Staining with anti-CK8/18 and anti-E-cadherin antibodies identified luminal epithelial cells. p63- and 34hE12-positive cells
depicted basal epithelial cells surrounding glandular structures. Prostate tumors often express neuroendocrine factors such as synaptophysin (Syn ). HPET and HPET-5
cells expressing synaptophysin were also observed. Bar, 50 Am. C, identification of cells that retain CD44 and Nestin expression. Both stem/progenitor cell markers
were expressed in the human tumor and HPET and HPET-5 tumors. Bar, 50 Am. Inserts highlight their staining pattern. Bar, 10 Am. D, diagrammatic summary of
the immunohistochemical analyses representing the differentiation of HPET and HPET-5 cells to luminal cells, basal cells, and neuroendocrine cells.

to determine whether a fraction of HPET and HPET-5 tumor cells

have growth potential similar to that of prostate stem cells. HPET
and HPET-5 tumors are regenerated through each cycle, reconstituting the original tumor and retain small subpopulations of cells
that express CD44 and Nestin (Fig. 6B), providing functional
evidence for the property of self-renewal.

Discussion
Similar to normal organs arising from normal stem cells, tumors
could be viewed as aberrant organs composed of heterogeneous
cell populations arising from cancer cells that have acquired the
capability for indefinite proliferation (1). Indeed, the clonal origin
of many tumors suggests that tumorigenic cells undergo processes
analogous to self-renewal and differentiation, generating self-

Cancer Res 2007; 67: (10). May 15, 2007

renewing cancer stem cell populations as well as cancer cells with

limited or no proliferative potential (1). The HPET clonal lines
support these observations in that they reconstitute the original
tumor, differentiate into basal, luminal, and neuroendocrine
epithelial cell lineages of the prostate (Fig. 4) and retain their
capacity for proliferation through serial transplantation (Fig. 6).
In the HPET and HPET-5 cell model, AR and p63 are absent until
glandular formation, when they are re-expressed in the appropriate
cell type in vivo. The prostate stem cell is thought to survive in an
androgen-deprived environment and does not express AR (30).
HPET and HPET-5 cells cultured in the absence or presence of
androgen do not express AR. In vivo, however, AR is expressed in
prostatic epithelium, suggesting that AR expression depends on
microenvironmental factors and correlates with cellular differentiation. AR has been associated with differentiation. Berger et al.

4812

www.aacrjournals.org

Prostate Cancer Cells with Stem Cell Characteristics

(31) showed that with the introduction of AR, HPE cell (PrEC) lines
immortalized with hTERT and transformed by SV40 large/small
T antigen (PrEC LHS-AR) or H-ras (PrEC LHSR-AR) underwent
partial differentiation into a luminal phenotype in vitro and in vivo.
In a different study, HPE cells immortalized with hTERT
differentiated into glandular buds in three-dimensional culture
on Matrigel (32). These structures consisted of a peripheral layer of
p63-positive cells surrounding luminal cells expressing low levels of
AR and PSA (32).
The origin of the prostate stem cell is still under debate. The p63
null mouse model (33, 34) suggests that epithelial development
does not occur in the absence of p63, which is highly expressed in
basal or progenitor layers of many epithelial tissues (35). However,
prostatic buds only appear on embryonic day 17.5 and are
underdeveloped at the time p63 / mice die perinatally. Signoretti
and Loda show that the whole prostatic epithelium is derived from
p63+ ROSA26 embryonic stem cells injected into p63 / blastocysts
in blastocyst complementation experiments (36). In contrast, using
renal grafting to rescue p63 / urogenital sinus and allow
differentiation, Kurita et al. show that in the absence of p63expressing basal cells, luminal secretory and neuroendocrine
epithelial cells still develop and are able to regenerate after
castration (37). In our study, clonal HPET-5 cells do not express

p63; yet, differentiation and glandular formation occur and p63 is

re-expressed in basal prostatic cells. A significant difference
between our study and the p63 / mouse model is that although
HPET p63 expression does not occur, the endogenous p63 gene is
still present. Under the appropriate stimuli in vivo, p63 protein is
re-expressed in the basal cell type.
Another basal cell marker is 34hE12. This marker is clinically
used to differentiate prostate cancer lesions from benign prostatic
glands. Although rare, 34hE12-positive cells have been detected in
prostatic adenocarcinoma (3840), indicating that prostate cancer
cells can exhibit characteristics attributed to a basal cell phenotype. Our study provides the first direct evidence that prostate
cancer stem/progenitor-like cells from single cell origin differentiate to luminal and neuroendocrine epithelial cells as well as into
basal cells in vivo, although these represent <0.1% of tumor cells.
In analyzing 34hE12 expression in HPET and HPET-5 tumors,
most if not all staining occurred where limited glandular
organization was still evident. The same pattern was observed
in the original human prostate cancer specimen. Alternatively,
although basal cells are primarily lost in adenocarcinoma,
prostate cancer stem cells likely have not lost the ability to
differentiate to p63- or 34hE12-positive cells. This may occur in
the appropriate microenvironment.

Figure 6. Histopathologic comparison of the original human tumor specimen with HPET and HPET-5 tumor grafts. A, a, original human tumor specimen. H&E
section of a radical prostatectomy whole-mount human specimen showing prostatic adenocarcinoma (Gleason grade 5 + 4 = 9) with neuroendocrine features. Within
the higher Gleason pattern, there are nuclei with "salt and pepper" chromatin and lack of prominent nucleoli suggestive of neuroendocrine differentiation. No
lymphovascular invasion is identified within these whole mount sections. b, human tumor containing Gleason pattern 4. c, HPET tumor graft. Prostatic adenocarcinoma
Gleason score 4 + 4 = 8 with abundant mitotic figures and apoptotic bodies. This tumor has abundant cribriforming patterns with salt and pepper nuclei similar to
the human tumor with a paucity of prominent nucleoli. There are focal areas of marked nuclear atypia that are also found in the human tumor. Sparse stroma is present
around glands with small, thin-walled vessels within stroma. d, small HPET focus containing Gleason pattern 5 (white arrow ). e, HPET-5 tumor graft. Prostatic
adenocarcinoma, Gleason grade 4 + 4 = 8 with neuroendocrine features, abundant mitotic figures and apoptosis. No Gleason pattern 5 is found. There is cytologic
atypia and sparse stroma around glands with small, thin-walled vessels within stroma. Bar , 50 Am. B, HPET and HPET-5 tumors are regenerated upon serial
transplantation. Serial transplantation was carried out by combining a small fragment of HPET or HPET-5 tumor tissue from the original graft with embryonic rUGM and
grafting them under the kidney capsule of SCID mice. Histologically, the regenerated tumors recapitulated the original tumors and retained small subpopulations of
cells expressing CD44 and Nestin. Bar , 50 Am.

www.aacrjournals.org

4813

Cancer Res 2007; 67: (10). May 15, 2007

Cancer Research

Histopathologic analyses indicate that HPET- and HPET-5

derived tumors reconstitute the Gleason score of the original
human tumor biopsy specimen in vivo. This model is unique in that
for the first time, pathologic evaluations and outcomes in response
to therapeutic treatment can be based on Gleason score. Gleason
score is a composite number composed of two cancer patterns
recognized by their architectural arrangement and is the most
frequently used grading system for human prostate cancer. The
combination of these patterns is a powerful prognostic indicator of
outcome as well as of biochemical failure for disease recurrence
(41). Gleason score also influences treatment and the ability to
predict the probability of lymph node metastasis (41). This scoring
system is not directly applicable to the transgenic mouse models
and is reviewed in detail elsewhere (42).
HPET and HPET-5 cells express the embryonic stem cell markers
Oct4, Nanog, and Sox2 and the early progenitor cell markers CD44,
Nestin, CD133, and c-kit. Transcription factors seem to act as
molecular switches that control cell fate during development. The
POU transcription factor Oct4 is expressed in the oocyte, declines
during the first two divisions, is re-expressed at the four- to eightcell stage and becomes restricted to the inner cell mass (ICM) of
the embryo (43). Oct4 maintains pluripotency, whereas upregulation of Oct4 expression results in differentiation to primitive
endoderm and mesoderm and down-regulation induces loss of
pluripotency and dedifferentiation into trophectoderm (22). Nanog
expression also occurs at the eight-cell morula stage and in ICM
(44). Increased Nanog expression prevents differentiation, and this
may act to restrict the differentiation-inducing potential of Oct4
(44, 45). Both Oct4 and Nanog expression is down-regulated as
embryogenesis progresses, and neither are expressed in normal
adult tissues (22, 44). Their role in maintaining HPET cell
pluripotency remains to be established.
A third member in this network of transcriptional factors is Sox2.
Sox2 is expressed with Oct4 and Nanog; however, it is not restricted
to multipotent embryonic cell lineages (46). Sox2 expression is also
present in developing brain, neural tube, sensory placodes, branchial
arches, and gut endoderm and continues to be expressed in adult

References
1. Reya T, Morrison SJ, Clarke MF, Weissman IL. Stem
cells, cancer, and cancer stem cells. Nature 2001;414:
10511.
2. Sell S, Pierce GB. Maturation arrest of stem cell
differentiation is a common pathway for the cellular
origin of teratocarcinomas and epithelial cancers. Lab
Invest 1994;70:622.
3. Roy NS, Nakano T, Keyoung HM, et al. Telomerase
immortalization of neuronally restricted progenitor cells
derived from the human fetal spinal cord. Nat
Biotechnol 2004;22:297305.
4. Wright WE, Piatyszek MA, Rainey WE, Byrd W, Shay
JW. Telomerase activity in human germline and
embryonic tissues and cells. Dev Genet 1996;18:1739.
5. Ostenfeld T, Caldwell MA, Prowse KR, Linskens MH,
Jauniaux E, Svendsen CN. Human neural precursor cells
express low levels of telomerase in vitro and show
diminishing cell proliferation with extensive axonal
outgrowth following transplantation. Exp Neurol 2000;
164:21526.
6. Yasunaga Y, Nakamura K, Ko D, et al. A novel human
cancer culture model for the study of prostate cancer.
Oncogene 2001;20:803641.
7. Yasunaga Y, Nakamura K, Ewing CM, Isaacs WB,
Hukku B, Rhim JS. A novel human cell culture model for

Cancer Res 2007; 67: (10). May 15, 2007

tissues such as the ovary (46). Interestingly, Sox2 seems to bind the
promoter of the Nestin gene to regulate its expression in neural
primordium (47). However, Nestin is not restricted to neural stem
cells, being expressed in embryonic stemderived progenitor cells
that can develop into neuroectodermal, endodermal, and mesodermal lineages (26). Prostate cancer stem cells seem to express
CD133 (30). CD133, initially isolated from neuroepithelial cells, is
also found in hematopoietic and endothelial progenitor cells (48). An
intriguing hypothesis is that neural crest is the source of all adult
stem cells in peripheral tissues (49), and that differentiation across
germ layers to many organs depends upon microenvironment (50).
Although HPET cells express Nestin and CD133, it is uncertain
whether prostatic HPET cells originate from neural crest.
The clonal HPET cancer stem/progenitor cell lines with
embryonic stem cell and early progenitor cell characteristics
provide the first in vivo model that recapitulates the original
Gleason score of the tumor from which they were derived. Thus,
this model may enhance our understanding of human tumor
development and provide a mechanism for preclinical testing of
diagnostic, treatment, and prevention strategies for cancer
patients. Furthermore, they provide cell lines critical to performing
in-depth mechanistic studies on the function of cancer stem cells
and epithelial cell differentiation. Because many epithelial cancers
seem to arise from cancer stem cells and often exhibit similar
characteristics, knowledge generated by the HPET cell lines will
likely be applicable to other epithelial cancers.

Acknowledgments
Received 12/21/2006; revised 3/6/2007; accepted 3/14/2007.
Grant support: National Institute of Diabetes and Digestive and Kidney Diseases
grants R01 DK60957 and R01 DK059142 and Frances Williams Preston Laboratories of
the T.J. Martell Foundation (S. Kasper).
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance
with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank Dr. Karin Williams for subcloning the EGFP gene into the Lentiviral
vector, Dr. Judith Campisi for providing the pBABE-PURO-hTERT vector, Anthony
Frazier for sectioning the human prostate whole mounts, and Erin Tillman for
proofreading the manuscript.

the study of familial prostate cancer. Cancer Res 2001;

61:596973.
8. Litvinov IV, Vander Griend DJ, Xu Y, Antony L,
Dalrymple SL, Isaacs JT. Low-calcium serum-free
defined medium selects for growth of normal prostatic
epithelial stem cells. Cancer Res 2006;66:8598607.
9. De Marzo AM, Meeker AK, Epstein JI, Coffey DS.
Prostate stem cell compartments: expression of the cell
cycle inhibitor p27Kip1 in normal, hyperplastic, and
neoplastic cells. Am J Pathol 1998;153:9119.
10. Tsujimura A, Koikawa Y, Salm S, et al. Proximal
location of mouse prostate epithelial stem cells: a model
of prostatic homeostasis. J Cell Biol 2002;157:125765.
11. Pitkanen-Arsiola T, Tillman JE, Gu G, et al. Androgen
and anti-androgen treatment modulates androgen
receptor activity and DJ-1 stability. Prostate 2006;66:
117793.
12. Hayward S, Cox S, Mitchell I, Hallowes R, Deshpande
N, Towler J. The effects of interferons on the activity of
alpha-glycerolphosphate dehydrogenase in benign prostatic hyperplasia cells in primary culture. J Urol 1987;
138:64853.
13. Kim NW, Piatyszek MA, Prowse KR, et al. Specific
association of human telomerase activity with immortal
cells and cancer. Science 1994;266:20115.
14. Bapat SA, Mali AM, Koppikar CB, Kurrey NK. Stem
and progenitor-like cells contribute to the aggressive

4814

behavior of human epithelial ovarian cancer. Cancer Res

2005;65:30259.
15. Levenberg S, Huang NF, Lavik E, Rogers AB, ItskovitzEldor J, Langer R. Differentiation of human embryonic
stem cells on three-dimensional polymer scaffolds. Proc
Natl Acad Sci U S A 2003;100:127416.
16. Cunha GR, Ricke W, Thomson A, et al. Hormonal,
cellular, and molecular regulation of normal and
neoplastic prostatic development. J Steroid Biochem
Mol Biol 2004;92:22136.
17. Huss WJ, Gray DR, Greenberg NM, Mohler JL, Smith
GJ. Breast cancer resistance protein-mediated efflux of
androgen in putative benign and malignant prostate
stem cells. Cancer Res 2005;65:664050.
18. Collins AT, Habib FK, Maitland NJ, Neal DE.
Identification and isolation of human prostate epithelial
stem cells based on alpha(2)beta(1)-integrin expression.
J Cell Sci 2001;114:386572.
19. Shackleton M, Vaillant F, Simpson KJ, et al.
Generation of a functional mammary gland from a
single stem cell. Nature 2006;439:848.
20. Boiani M, Scholer HR. Regulatory networks in
embryo-derived pluripotent stem cells. Nat Rev Mol
Cell Biol 2005;6:87284.
21. Boyer LA, Lee TI, Cole MF, et al. Core transcriptional
regulatory circuitry in human embryonic stem cells. Cell
2005;122:94756.

www.aacrjournals.org

Prostate Cancer Cells with Stem Cell Characteristics

22. Niwa H, Miyazaki J, Smith AG. Quantitative expression of Oct-3/4 defines differentiation, dedifferentiation
or self-renewal of ES cells. Nat Genet 2000;24:3726.
23. Bhatia B, Tang S, Yang P, et al. Cell-autonomous
induction of functional tumor suppressor 15-lipoxygenase 2 (15-LOX2) contributes to replicative senescence
of human prostate progenitor cells. Oncogene 2005;24:
358395.
24. Kania G, Corbeil D, Fuchs J, et al. Somatic stem cell
marker prominin-1/CD133 is expressed in embryonic stem
cell-derived progenitors. Stem Cells 2005;23:791804.
25. Collins AT, Berry PA, Hyde C, Stower MJ, Maitland NJ.
Prospective identification of tumorigenic prostate
cancer stem cells. Cancer Res 2005;65:1094651.
26. Wiese C, Rolletschek A, Kania G, et al. Nestin
expression: a property of multi-lineage progenitor cells?
Cell Mol Life Sci 2004;61:251022.
27. Ronnstrand L. Signal transduction via the stem cell
factor receptor/c-Kit. Cell Mol Life Sci 2004;61:253548.
28. Reubinoff BE, Itsykson P, Turetsky T, et al. Neural
progenitors from human embryonic stem cells. Nat
Biotechnol 2001;19:113440.
29. Heinlein CA, Chang C. Androgen receptor in prostate
cancer. Endocr Rev 2004;25:276308.
30. Rizzo S, Attard G, Hudson DL. Prostate epithelial
stem cells. Cell Prolif 2005;38:36374.
31. Berger R, Febbo PG, Majumder PK, et al. Androgeninduced differentiation and tumorigenicity of human
prostate epithelial cells. Cancer Res 2004;64:886775.
32. Kogan I, Goldfinger N, Milyavsky M, et al. hTERT-

www.aacrjournals.org

immortalized prostate epithelial and stromal-derived

cells: an authentic in vitro model for differentiation and
carcinogenesis. Cancer Res 2006;66:353140.
33. Mills AA, Zheng B, Wang XJ, Vogel H, Roop DR,
Bradley A. p63 is a p53 homologue required for limb and
epidermal morphogenesis. Nature 1999;398:70813.
34. Yang A, Schweitzer R, Sun D, et al. p63 is essential for
regenerative proliferation in limb, craniofacial and
epithelial development. Nature 1999;398:7148.
35. Barbieri CE, Pietenpol JA. p63 and epithelial biology.
Exp Cell Res 2006;312:695706.
36. Signoretti S, Loda M. Defining cell lineages in the
prostate epithelium. Cell Cycle 2006;5:13841.
37. Kurita T, Medina RT, Mills AA, Cunha GR. Role of p63
and basal cells in the prostate. Development 2004;131:
495564.
38. Googe PB, McGinley KM, Fitzgibbon JF. Anticytokeratin antibody 34 beta E12 staining in prostate carcinoma. Am J Clin Pathol 1997;107:21923.
39. Oliai BR, Kahane H, Epstein JI. Can basal cells be
seen in adenocarcinoma of the prostate?: an immunohistochemical study using high molecular weight
cytokeratin (clone 34betaE12) antibody. Am J Surg
Pathol 2002;26:115160.
40. Yang XJ, Lecksell K, Gaudin P, Epstein JI. Rare
expression of high-molecular-weight cytokeratin in
adenocarcinoma of the prostate gland: a study of 100
cases of metastatic and locally advanced prostate
cancer. Am J Surg Pathol 1999;23:14752.
41. DeMarzo AM, Nelson WG, Isaacs WB, Epstein JI.

4815

Pathological and molecular aspects of prostate cancer.

Lancet 2003;361:95564.
42. Shappell SB, Roberts RL, Thomas GV, et al. Prostate
pathology of genetically engineered mice: the consensus
report from the Bar Harbor meeting. Cancer Res 2004;
64:2270305.
43. Stewart CL. Oct-4, scene 1: the drama of mouse
development. Nat Genet 2000;24:32830.
44. Chambers I, Colby D, Robertson M, et al. Functional
expression cloning of Nanog, a pluripotency sustaining
factor in embryonic stem cells. Cell 2003;113:64355.
45. Mitsui K, Tokuzawa Y, Itoh H, et al. The homeoprotein
Nanog is required for maintenance of pluripotency in
mouse epiblast and ES cells. Cell 2003;113:63142.
46. Avilion AA, Nicolis SK, Pevny LH, Perez L, Vivian N,
Lovell-Badge R. Multipotent cell lineages in early mouse
development depend on SOX2 function. Genes Dev
2003;17:12640.
47. Tanaka S, Kamachi Y, Tanouchi A, Hamada H, Jing N,
Kondoh H. Interplay of SOX and POU factors in
regulation of the Nestin gene in neural primordial cells.
Mol Cell Biol 2004;24:883446.
48. Shmelkov SV, St Clair R, Lyden D, Rafii S. AC133/CD133/
Prominin-1. Int J Biochem Cell Biol 2005;37:7159.
49. Pierret C, Spears K, Maruniak JA, Kirk MD. Neural
crest as the source of adult stem cells. Stem Cells Dev
2006;15:28691.
50. Vescovi A, Gritti A, Cossu G, Galli R. Neural stem
cells: plasticity and their transdifferentiation potential.
Cells Tissues Organs 2002;171:6476.

Cancer Res 2007; 67: (10). May 15, 2007